A new study looking at the efficacy of closed-loop pump systems was recently published in The New England Journal of Medicine. Similar to other studies that have compared data of participants on closed-loop pump therapy to those on sensor-augmented or traditional pump therapy, these researchers found a significant increase in time spent in range and an improvement in A1C for the group using the closed-loop system.

Unlike other recent studies, this project was more substantial in that it followed participants for a full six months and used newer tech with the power to correct blood sugar levels beyond basal rate adjustment.

While their findings should bring hope to the diabetic community as a whole, it is important to dig deeper into the limitations of this new technology. Closed-loop systems bring the promise of an “artificial pancreas” but often fall short in delivering those goals. This seems clearly outlined in the less celebrated data presented in this study.

How the Study Was Structured

This trial enrolled 168 Type 1 diabetic patients between the ages of 14 and 71. The subjects varied in both how they managed their diabetes prior to the study — some used MDI, some pumps, and some individuals from each group used CGMs — and their starting A1C, which ranged from 5.4 to 10.6.

The participants were randomly assigned to one of two groups. The control group was given sensor-augmented pumps without any automated adjustment functions such as suspension of insulin delivery for low readings. The other group was given the yet to be FDA approved Tandem t:slim pump with Control-IQ technology paired with a Dexcom CGM.

All participants were given training on their assigned system as well as a two to eight-week “run-in” phase to allow participants to get used to the new equipment as needed. Base data was also collected from each patient at the beginning of the study period.

Additional data were collected at pre-scheduled intervals over the course of six months. Information was uploaded from CGMs, pumps, and meters and blood testing was done to check A1C levels and blood ketones. Each subject was also asked to report any adverse events such as severe low blood sugar or positive ketone tests.

Study Results

Researchers found that the group utilizing the closed-loop therapy had an 11% increase in time-in-range compared to baseline, while the control group saw no change. While this number may sound somewhat insignificant, it translates to an average of 2.6 hours per day less time with high blood sugar (over 180 mg/dl). Low blood sugars were also decreased by about 13 minutes per day.

The closed-loop system was especially helpful in maintaining healthy blood sugars during the latter part of the night into the early morning, allowing patients to wake more often with near-normal blood sugars.

Researchers also saw a decrease in A1C in the closed-loop group while the control group remained about the same.

These benefits were seen within the first month of the trial and continued throughout the six-month study.

Looking Beyond the Numbers

This study highlights the positive effect that more advanced devices can have on diabetes management. When comparing traditional DMI therapy, pump therapy, and the use of CGMs, it is clear that the most effective treatment for increasing time-in-range and lowering A1C is a closed-loop system that reduces the risk of human error.

However, it is important that we recognize the limitations of current technology.

The study manuscript published in NEJM referred to the Tandem Control-IQ system as a “closed-loop” pump, even going so far as to call it an “artificial pancreas.” While it is true that the Tandem pump has more capabilities than the current front-runner in looping technology, the Medtronic MiniMed Hybrid Closed-Loop, neither, in my opinion, comes close to earning the title of an artificial pancreas.

Like the Medtronic pump system, this Tandem pump makes adjustments to basal insulin rates based on changes in blood sugar reported by the CGM. 

What sets Tandem apart is its ability to give automatic correction boluses for spiking blood sugars, a feature meant to reduce the impact of a missed or forgotten meal boluses. 

Both Medtronic and Tandem systems still require the user to calculate the carbohydrate content of their meals and bolus for them manually, which remains an open window for human error. 

The Tandem system, however, corrects automatically if you are wrong.

According to Tandem’s info, this new system can give correction boluses up to 60% of the bolus that was missed. This auto-correction should eventually bring your blood sugar back down to normal but your blood glucose would likely spike at a fairly high number due to the smaller late bolus and then drop and level out at higher than average. For example, with a carb calculated bolus, it might take two hours after a meal to get back in range but with auto-correction alone, it might take an hour or two longer to get back in range.

Additionally, the number of adverse events recorded during this study showed surprisingly more occurring in the closed-loop group than in the control. The majority of these events were various degrees of ketosis. In fairness, most occurred due to infusion set failure and not because of a flaw in the algorithm or the pump system itself.

Opinion

The persistence of ketone-inducing hyperglycemia even when utilizing such a high-tech piece of equipment is frustrating. It serves to highlight one of the often overlooked aspects of increasingly intelligent diabetes tech: the blind reliance on these devices to keep us safe and healthy.

The more advanced diabetes devices become, the easier it will be to put them in the driver’s seat while taking a backseat to management ourselves. So long as there are flaws in the way these systems work and inherent issues in how they deliver (or not) insulin to the patient, these adverse events are only likely to increase, even as overall time-out-of-range decreases.

Without a doubt, options like hybrid closed-loop systems are helping improve diabetes outcomes worldwide. But, so long as these systems leave room for human error or retain the potential for function interruption, I believe the distinction of “artificial pancreas” should be withheld.